In recent years, much research has been conducted on path replanning in a dynamic environment, as it is considered to be a more complex problem than path planning in a static environment. However, with respect to determining when to replan a path, most research on dynamic path replanning typically generates a new path if obstacles collide with the current path, or with reference to each time unit interval. As these approaches do not focus on effective path regeneration, they result in redundant and unnecessary path regeneration. In this context, this paper proposes an intelligent decision-making mechanism for path regeneration to determine whether a new path should be generated when obstacles move over time. Based on the identification of certain important obstacles, we let these obstacles propagate influence, and then obtain path values that represent the distance between the obstacles and the path. Then, the necessity of path regeneration is determined according to the changes in path values in a dynamic environment. Using this approach, it is possible to both enhance paths in terms of distance and resolve any infeasibilities due to moving obstacles. In the experiment, we test the proposed methodology to verify that appropriate decisions are made, and compare it with other approaches that regenerate a path at each time interval or on collision. Our results show that the paths are regenerated only if it is necessary to improve the current path or to resolve infeasibilities, indicating that effective and intelligent decisions for path regeneration can be obtained using the proposed decision mechanism.
Bibliographical noteFunding Information:
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (No. 2016R1D1A1B03935266 ).
© 2018 Elsevier Inc.
- Distance propagation
- Dynamic environment
- Path replanning
ASJC Scopus subject areas
- Control and Systems Engineering
- Theoretical Computer Science
- Computer Science Applications
- Information Systems and Management
- Artificial Intelligence